CIFS/SMB3 Kernel Maintainer Steve French is Dead

For nearly a quarter century, one name has sat atop every commit touching Linux's CIFS and SMB3 code: Steve French. That changes with Linux 7.3. French has stepped down as maintainer due to health reasons, closing out one of the longest continuous maintainership runs in the kernel's history.

If you have ever mounted a Windows share on a Linux box, pulled files from a NAS over SMB3, or run Samba in a mixed-OS office, you have relied on code French wrote or reviewed. He did not just maintain the subsystem — he built most of it from scratch. He then spent years hardening it while working at Microsoft, a company whose products depend on the very protocol he was upstreaming into a competing operating system.

Context

CIFS, which developers later folded into the broader SMB3 implementation, is the kernel-side client that lets Linux talk to Windows file shares, NAS boxes, and Samba servers using Microsoft's native file-sharing protocol. Before this code existed in a usable form, getting Linux and Windows machines to share files reliably meant leaning entirely on Samba running in userspace or accepting janky, unencrypted legacy CIFS support.

French started that work in the early 2000s and never really let go of it. He pushed the driver through SMB1's aging feature set into SMB2 and SMB3, adding support for encryption, multichannel connections, directory leases, and the kind of Windows Active Directory integration that enterprise IT departments actually need. That's not a small technical lift. SMB3 is a stateful, session-oriented protocol with plenty of Windows-specific quirks, and getting it right inside a kernel module means constant testing against real Windows Server behavior, not just the written spec.

What made French's tenure unusual was his day job. Microsoft employs him, and Microsoft owns the protocol he was implementing in a rival kernel. Instead of creating a conflict, that arrangement turned into one of the more productive stories in open source infrastructure.

French's institutional knowledge of SMB3's internals, gained from the vendor side, is not something an outside contributor could easily replicate. He used that access to keep the Linux client compatible as Windows Server and Azure Files evolved their SMB implementations. He did this while also serving as a longtime Samba team member and a familiar face at kernel and file systems conferences. See related: eevdf vs cfs: what really changed in linux scheduling for additional background.

A health-related maintainer transition is not new territory for the kernel project. Long-running subsystems eventually outlive the person who started them, and the kernel community has handled these handoffs before — sometimes smoothly, sometimes with a rocky adjustment period while a new maintainer learns the accumulated history of a codebase. CIFS/SMB3 carries a particularly dense set of that history, spanning encryption negotiation, POSIX extensions Microsoft added for Linux interoperability, and years of edge-case bug fixes tied to specific Windows Server patch levels.

Implications

For everyday desktop users mounting a shared drive, nothing changes immediately. The code French leaves behind is mature and well-tested, and a maintainer transition does not erase years of stability work. Your mount.cifs commands, your /etc/fstab entries pointing at a Windows share, and your Samba-based home NAS setups will keep working exactly as they did.

Also read: is kde plasma 6.8 worth upgrading for performance?: the details

The real question is pace and depth of future development. New SMB3 features, performance tuning for multichannel setups, and fast turnaround on interoperability bugs against new Windows Server releases have all depended on French's direct line into Microsoft's own SMB engineering. Whoever picks up the maintainer role, whether an existing co-maintainer or a new name entirely, will need to either inherit that access or build equivalent relationships from scratch — and that is not a quick substitution.

For Linux system administrators running mixed Windows-Linux environments, this is worth flagging as a risk to monitor rather than an emergency. If you manage file servers where Linux clients authenticate against Active Directory and mount SMB3 shares for daily operations, keep an eye on the CIFS/SMB3 mailing list and kernel changelogs over the next few kernel cycles. Watch specifically for how quickly maintainers triage reported regressions and whether feature parity with new Windows Server SMB dialects continues at the same clip.

There is also a broader lesson here for how the kernel handles vendor-adjacent maintainers. French's dual role as a Microsoft employee and the person responsible for Linux's SMB3 client was, for years, one of the quieter success stories of cross-company open source collaboration. Microsoft should keep investing in that arrangement, and the kernel community should watch closely, so that something equally credible replaces it rather than letting it quietly degrade.

Watch the kernel mailing list threads and the next few merge windows for the name attached to CIFS/SMB3 pull requests. That will tell you faster than any announcement whether the transition is a smooth handoff of deep protocol knowledge or the start of a slower, bumpier period for one of Linux's most quietly essential file-sharing subsystems.